Heavy drinking during the college years and in the early twenties remains a significant risk factor for adjustment problems, accidents, and later addiction. Young people, however, also show high rates of remission of alcohol problems in longitudinal research. Longitudinal studies of adolescent populations suggest that adult role demands are associated with remission of drinking problems. Recent studies of children of alcoholics suggest that less subjective intoxication and greater stress response dampening when alcohol is consumed may be related to risk of addiction. The current proposal will combine laboratory and survey methods in a study of high risk drinkers in order to predict chronicity of heavy drinking. Selected biological, psychological, and social network variables will be assessed in order to predict continuity of heavy drinking and natural moderation process after graduation from college. Better understanding of natural change and moderate drinking practices, and facilitate the early detection or prediction of the chronicity of heavy drinking. Heavy drinking (less than 15 drinks per week) college seniors (n = 300) will be recruited for a three-year longitudinal study of the prediction of future drinking habits. Subjects will be assessed initially during the academic quarter prior to graduation from college. Assessment will include measures of drinking patterns and history, family history of alcoholism, impulsivity, type of living setting, social network alcohol consumption, perceived alcohol effects and perceived norms. A laboratory alcohol challenge will assess subjective intoxication and stress response dampening. Annual assessments over the next three years will allow for the multivariate prospective prediction of changes in drinking habits. It is hypothesized that less subjective intoxication and greater stress response dampening will be associated with continuity of heavy drinking. Changes in drinking patterns should also be accompanied by lifestyle changes and reductions in drinking among social contacts, as well as changes in beliefs about alcohol effects and normative drinking. Prospective analyses will examine if changes in drinking result from or predate changes in social and occupational functioning and alcohol beliefs.